Landcover Classification

A Computer Vision Model For Classifying Terrain.
⭐️ As Featured on Sonoma.edu
🎥 Watch the project video

Table of Contents

1. About The Project
2. Getting Started
   • Running The Program
3. Roadmap
4. Usage
5. Contributors
6. Connect
7. License
8. Acknowledgments

About The Project

This project was completed as a senior capstone project at Sonoma State University in fall 2024.

A Larger Initiative: BioSCape

This project is part of the larger BioSCape initiative, a collaboration aimed at understanding biodiversity in the Greater Cape Floristic Region.

The Biodiversity Survey of the Cape is a NASA-SANSA Biodiversity research project focused on the Greater Cape Floristic Region of South Africa.

• This initiative utilizes hyperspectral images captured by the AVIRIS-NG remote sensor to analyze and classify land cover types. These hyperspectral images, containing 432 bands, provide detailed spectral data that enable precise classification across diverse ecological categories.

Learn more about BioSCape.

Land Cover Classification Categories

• Wetlands
• Planted Forest
• Permanent Crops (e.g., vineyard)
• Unconsolidated Barren
• Natural Grassland
• Consolidated Barren (e.g., rocks, salt pans)
• Built-up Areas
• Mixed or Not Classified
• Natural Wooded Land
• Waterbodies
• Annual Crops (e.g., wheat)
• Shrubs

Objectives

1. Hyperspectral Land Cover Mapping
   Leveraged hyperspectral imaging from the remote sensor AVIRIS-NG to classify the terrain into the categories listed above.

2. 2D Classification Approach
   Explored a two-dimensional (2D) methodology for land cover classification. See the 1D approach here.

Getting Started

Running The Program

1. Clone the repo

git clone https://github.com/V1sionaries-Landcover-Classification.git

Roadmap

See the open issues for a full list of proposed features (and known issues).

Contributors

Four students worked together to implement this project.

Connect With Us

Blake Marshall - LinkedIn
Sean Farmer - LinkedIn
Jacob Sellers - LinkedIn
Isauro Ramos - LinkedIn

License

Distributed under the MIT License. See LICENSE.txt for more information.

(back to top)